Classification of Customer Satisfaction in Marketplace

Abstract

Customer satisfaction is one of the most important things in the marketplace to evaluate the customer’s impression of the service in the marketplace. Customer satisfaction is measured in five aspects, i.e., the application user interface, application features, discounts and promotions, ease of access to the applications, and customer service. This paper used three classification algorithms in data mining: KNN, Decision Tree, and Naïve Bayes to classify customer satisfaction. Calculation of the algorithms performed in this paper using the Rapid Miner. There are two experiments on each algorithm using the RapidMiner application, but the method used to calculate data is undoubtedly different. There are 549 data taken from marketplace user surveys, namely Shopee, Tokopedia, and Lazada. The highest accuracy result obtained was 96.36%, with a precision of 96.97% and a recall of 98.96% of the classification using the Decision Tree algorithm through the second experiment. From these results, it can be inferred that Decision Tree is the most accurate algorithm in this paper because it has the highest accuracy when performing experiments. According to the results of this research, we desire to be an accurate reference source that can be used by marketplaces to obtain information on customer satisfaction and improve their service.